Article Text
Statistics from Altmetric.com
Cervical artery dissection (CAD) accounts for 2% of all ischaemic strokes.1 With improved non-invasive neurovascular imaging methods it is now appreciated that carotid and vertebral dissections are much more prevalent than previously believed and account for 25–30% of all ischaemic strokes in young patients aged <50 years. Despite their importance, best therapeutic management remains unclear and is not evidence based.
There are no randomised trials of any treatment regimen in cervical dissection. Following presentation with either local symptoms (pain, headache or Horner’s syndrome) or stroke/TIA there is a risk of stroke, particularly in the first month.2 3 As this is probably thromboembolic, anticoagulation for 3–6 months is often recommended. However, this is not evidence based and there are no data from randomised controlled trials.4 Previous surveys to determine choice of antithrombotic agent in secondary prevention of stroke have shown a bias towards anticoagulants (up to 81% in one Canadian survey).5
Methodology
A postal questionnaire was distributed to members of the British Association of Stroke Physicians (which is the professional body for consultant neurologists, geriatricians and general physicians specialising in stroke). Clinicians were asked the following questions: How many cervical dissections do you see per year? What is your preferred method of imaging? What is your preferred treatment? Would you thrombolyse patients with cervical artery dissection? Treatment of intracranial dissection was excluded in view of the hazards of subarachnoid haemorrhage.
Results
From a total of 330 questionnaires, 196 (59%) physicians responded, of whom 23 were neurologists, 167 were stroke physicians/geriatricians and six were consultants in other branches of internal medicine. The number of patients seen per year with dissection varied widely (table 1). Most (70%) physicians saw between one and five patients per year.
Magnetic resonance (MRI and/or MRA) was the most widely used diagnostic imaging procedure (90%). CT angiography, digital subtraction angiography and carotid ultrasound were less frequently relied upon. Most physicians (65%) chose anticoagulants as the drug of choice (50% anticoagulation always and 15% anticoagulants or antiplatelet agents). However, a large minority (30%) always used antiplatelet agents. Eight per cent were prepared to use thrombolysis in patients with cervical artery dissection. Stenting was hardly used (1%) for treatment of dissection.
Conclusion
This questionnaire demonstrates marked variability in treatment administered to patients with cervical dissection. While half of respondents would always anticoagulate, 30% always gave antiplatelet agents. These differing views reflect the lack of evidence on which to base a therapeutic decision and emphasise the need for a randomised controlled trial comparing antiplatelet agents with anticoagulation for acute cervical dissection. There has been concern that clinicians would not be prepared to randomise to antiplatelet agents, but the variability in treatments used in routine practice suggests this may not be the case. The relatively low published annual event rates of stroke and death in CAD have dissuaded previous investigators from undertaking a randomised therapeutic trial because of the large numbers of patients needed. Based on retrospective uncontrolled data, the Cochrane review concluded a total of 2800 patients would be needed in a two arm study.4 More recent data suggest the 1 year stroke, death and transient ischaemic attack rate after presentation is >10%, suggesting sample size estimates may be lower.2 With such a low incidence of cervical artery dissection, such a trial although challenging, would be possible.
The Cervical Artery Dissection In Stroke Study (CADISS) (www.dissection.co.uk) has recently been started to address this question. This is a multicentre randomised prospective open label study comparing antiplatelet to anticoagulant therapy with heparin and then warfarin. A feasibility phase, funded by the Stroke Association, has begun. This will enrol 250 patients and allow feasibility of recruitment to be evaluated, with primary endpoints of stroke and death at 3 months. Transient ischaemic attacks were considered only secondary end points in view of the possibility of potential bias. This feasibility phase could be a prelude to a large international study with sample size calculations based on the unblinded end point frequency identified during the feasibility phase.
Our questionnaire reflects only the management practices of physicians in the UK. However, from discussion with colleagues worldwide there appears to be widespread uncertainty as to the best antithrombotic therapy in this condition. It is possible the reluctance to use thrombolysis may reflect the lower use of this treatment for stroke in general in the UK.
Footnotes
Competing interests: None.